Overview
Hands-on lead who builds production AI agent platforms for clients on the cloud, turning designs into fast, reliable, secure and cost-efficient working platforms, and leads a team of engineers working directly with client technical teams.
What you'll do
- Build the core parts of an agent platform: agent orchestration, connections to client tools and data using the Model Context Protocol, a model gateway with routing and fallback, short and long-term memory, and tests that check agent quality
- Deploy agents on managed cloud agent services and run them securely with the right identity and permissions
- Build the agent harness: the layer that handles tool calls, context, approvals, retries and session state
- Set up infrastructure as code, CI/CD and GitOps for agent releases
- Make the platform handle high volume: autoscaling, queues, caching, rate-limit handling, and switching to another model or region when one is slow or down
- Run load and performance tests, find bottlenecks and fix them
- Set up tracing, metrics and cost tracking for every agent and tool call
- Study the client's current setup, identify quick wins and deliver them in the first weeks
- Keep client teams updated on progress and risks
- Review code and designs, coach and lead engineers
- Build agents for custom/legacy technology like Java and Mainframe, and platforms like SAP, Oracle and Workday
What you'll need
- Minimum 12 years of experience
- 15 years full time education
- Must have skill: Automation in Application Development
- Strong in one cloud, working knowledge of another: AWS (Bedrock AgentCore, Bedrock, EKS, Lambda), Azure (Microsoft Foundry Agent Service, Azure OpenAI, AKS, Entra ID), or GCP (Gemini Enterprise Agent Platform, Cloud Run, GKE)
- Strong Python (async programming, FastAPI, APIs), minimum 3 years of Python experience
- Containers, Kubernetes, Terraform or CDK
- CI/CD: GitHub Actions, Azure DevOps, Jenkins or similar
- Messaging and caching: Kafka, SQS, Pub/Sub, Service Bus
- OpenTelemetry, Prometheus, Grafana, cloud-native monitoring
- Security basics: access control, secrets, private networking
Details
- Based at the Bangalore office
Read the full description and apply on the company’s own careers page.